The primary objective of this paper is to measure persistence in overall inflation measured by WPI and CPI-IW and their components for the Indian economy over the period 1996:4 to 2017:02. The study first tests for persistence and then quantifies it using various univariate and multivariate approach. The univariate approach estimates persistence using sum of autoregressive coefficient, largest root of the autoregressive process, spectral measures such as Campbell and Mankiw measure of persistence and spectral density function at frequency zero. Estimates of persistence are also derived using variance ratio tests and rolling regressions technique. The study also uses long memory approach (ARFIMA) to measure persistence. In a multivariate approach, the study examines structural form of persistence by employing a time varying vector autoregressive model (TVP-VAR) with stochastic volatility, incorporating inflation, interest rate, demand and supply side variables as its constituents. Results of the study suggest that persistence is high in the economy at both aggregate and disaggregate level. Wholesale price inflation shows intermediate memory, however Consumer prices exhibit long-memory in a univariate analysis. However when structural persistence is studied then WPI inflation is affected by exchange rate shocks and supply shocks such as fuel but it is found to be invariant with respect to interest rate shocks. On the other hand, CPI-IW inflation exhibits time variation with respect to food shocks and exchange rate shocks but is found to be mostly invariant with respect to interest rate shocks.

本文的主要目标是衡量 1996:4 至 2017:02 期间由 WPI 和 CPI-IW 及其组成部分衡量的印度经济总体通胀的持续性。该研究首先测试持久性，然后使用各种单变量和多变量方法对其进行量化。单变量方法使用自回归系数的总和、自回归过程的最大根、谱度量（如 Campbell 和 Mankiw 持久性度量和频率为零的谱密度函数）之和来估计持久性。还使用方差比检验和滚动回归技术得出持久性的估计值。该研究还使用长记忆法 (ARFIMA) 来衡量持久性。在多元方法中，该研究通过采用具有随机波动性的时变向量自回归模型 (TVP-VAR) 来检查持久性的结构形式，将通货膨胀、利率、需求和供应侧变量作为其组成部分。研究结果表明，经济在总体和分解层面的持久性都很高。批发价格通胀表现出中间记忆，但消费者价格在单变量分析中表现出长期记忆。然而，当研究结构持久性时，WPI 通胀受汇率冲击和供应冲击（如燃料）的影响，但发现它对利率冲击是不变的。另一方面，